TagRec

Abstract

This article presents TagRec , a framework to foster reproducible evaluation and development of recommender algorithms based on folksonomy data. The purpose of TagRec is to provide the research community with a standardised framework that supports all steps of the development process and the evaluation of tag-based recommendation algorithms in a reproducible way, including methods for data pre-processing, data modeling and recommender evaluation. TagRec currently contains 32 state-of-the-art algorithms for tag and item prediction, including a set of novel and very efficient algorithms based on the human cognition theories ACT-R and MINERVA2. The framework should be relevant for researchers, teachers, students and developers working on recommender systems and predictive modeling in general and those interested in tag-based recommender algorithms in particular.